diff --git a/include/boost/utility/value_init.hpp b/include/boost/utility/value_init.hpp
index c4cd3e1..f5f04bf 100644
--- a/include/boost/utility/value_init.hpp
+++ b/include/boost/utility/value_init.hpp
@@ -7,103 +7,106 @@
 // 21 Ago 2002 (Created) Fernando Cacciola
 // 07 Set 2007 (Worked around MSVC++ bug) Fernando Cacciola, Niels Dekker
 // 16 Nov 2007 (Refactoring: removed private base classes) Fernando Cacciola, Niels Dekker
+// 09 Dec 2007 (Worked around various compiler bugs) Fernando Cacciola, Niels Dekker
 //
 #ifndef BOOST_UTILITY_VALUE_INIT_21AGO2002_HPP
 #define BOOST_UTILITY_VALUE_INIT_21AGO2002_HPP
 
-#include <boost/detail/select_type.hpp>
-#include <boost/type_traits/cv_traits.hpp>
-#include <boost/detail/workaround.hpp>
-
-// Microsoft Visual C++ does not correctly support value initialization, as reported by
-// Pavel Kuznetsov (MetaCommunications Engineering), 7/28/2005, Feedback ID 100744,
-// Feedback Title: Value-initialization in new-expression
+// Note: The implementation of boost::value_initialized had to deal with the
+// fact that various compilers haven't fully implemented value-initialization:
+// Microsoft Feedback ID 100744 - Value-initialization in new-expression
+// Reported by Pavel Kuznetsov (MetaCommunications Engineering), 2005-07-28
 // https://connect.microsoft.com/VisualStudio/feedback/ViewFeedback.aspx?FeedbackID=100744
-// The report was closed at 11/14/2006, and its status was set to "Closed (Won't Fix)".
-// Luckily, even in the presence of this compiler bug, boost::value_initialized will still
-// do its job correctly, when using the following workaround:
-#if BOOST_WORKAROUND(BOOST_MSVC, BOOST_TESTED_AT(1500))
-#  define BOOST_UTILITY_VALUE_INIT_WORKAROUND
-#endif
-
-#ifdef BOOST_UTILITY_VALUE_INIT_WORKAROUND
+// GCC Bug 30111 - Value-initialization of POD base class doesn't initialize members
+// Reported by Jonathan Wakely, 2006-12-07 
+// http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30111
+// GCC Bug 33916 - Default constructor fails to initialize array members
+// Reported by Michael Elizabeth Chastain, 2007-10-26
+// http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33916
+// Borland Report 51854 - Value-initialization: POD struct should be zero-initialized
+// Reported by Niels Dekker (LKEB, Leiden University Medical Center), 2007-11-09
+// http://qc.codegear.com/wc/qcmain.aspx?d=51854
+// The constructor of boost::value_initialized<T> works around these issues, by
+// clearing the bytes of T, before constructing the T object it contains. 
 
 #include <boost/aligned_storage.hpp>
+#include <boost/detail/select_type.hpp>
+#include <boost/detail/workaround.hpp>
+#include <boost/type_traits/cv_traits.hpp>
 #include <boost/type_traits/alignment_of.hpp>
 #include <cstring>
 #include <new>
 
-#ifdef BOOST_MSVC
-#pragma warning(push)
-#if _MSC_VER >= 1310
-// When using MSVC 7.1 or higher, placement new, "new (&x) T()", may trigger warning C4345:
-// "behavior change: an object of POD type constructed with an initializer of the form ()
-// will be default-initialized".  There is no need to worry about this, though.
-#pragma warning(disable: 4345)
-#endif
-#endif
-
 namespace boost {
 
 template<class T>
 class value_initialized
 {
   private :
-    mutable typename ::boost::aligned_storage<sizeof(T), ::boost::alignment_of<T>::value>::type x;
+    struct wrapper
+    {
+#if !BOOST_WORKAROUND(__BORLANDC__, BOOST_TESTED_AT(0x592))
+      typename
+#endif 
+      remove_const<T>::type data;
+    };
+
+    mutable
+#if !BOOST_WORKAROUND(__BORLANDC__, BOOST_TESTED_AT(0x592))
+      typename
+#endif
+      ::boost::aligned_storage<sizeof(wrapper), ::boost::alignment_of<wrapper>::value>::type x;
 
   public :
 
     value_initialized()
     {
       std::memset(&x, 0, sizeof(x));
-      new (&x) T();
+#ifdef BOOST_MSVC
+#pragma warning(push)
+#if _MSC_VER >= 1310
+// When using MSVC 7.1 or higher, the following placement new expression may trigger warning C4345:
+// "behavior change: an object of POD type constructed with an initializer of the form ()
+// will be default-initialized".  There is no need to worry about this, though.
+#pragma warning(disable: 4345)
+#endif
+#endif
+      new (&x) wrapper();
+#ifdef BOOST_MSVC
+#pragma warning(pop)
+#endif
+    }
+
+    value_initialized(value_initialized const & arg)
+    {
+      void const * const ptr = &(arg.x);
+      new (&x) wrapper( *static_cast<wrapper const *>(ptr) );
+    }
+
+    value_initialized & operator=(value_initialized const & arg)
+    {
+      T & this_data = this->data();
+      T const & arg_data = arg.data();
+      this_data = arg_data;
+      return *this;
     }
 
     ~value_initialized()
     {
-      void * ptr = &x; 
-      static_cast<T*>(ptr)->T::~T(); 
+      void * const ptr = &x; 
+      static_cast<wrapper *>(ptr)->wrapper::~wrapper();
     }
 
     T& data() const
     {
-      void * ptr = &x;
-      return *static_cast<T*>(ptr);
+      void * const ptr = &x;
+      return static_cast<wrapper *>(ptr)->data;
     }
 
     operator T&() const { return this->data(); }
 
 } ;
 
-#ifdef BOOST_MSVC
-// Restores the state of warning C4345.
-#pragma warning(pop)
-#endif
-
-#else
-
-namespace boost {
-
-template<class T>
-class value_initialized
-{
-  public :
-
-    value_initialized() : x() {}
-
-    T& data() const { return x; }
-
-    operator T&() const { return this->data(); }
-
-    mutable
-#if !BOOST_WORKAROUND(__BORLANDC__, BOOST_TESTED_AT(0x564))
-    typename 
-#endif
-    ::boost::remove_const<T>::type x ;
-
-} ;
-#endif
-
 
 
 template<class T>